Association between Greenness Structures and Frailty in an Elderly Prospective Longitudinal Cohort in China

28 Background: Frailty is the accumulation of aging-induced deficits, leading to vulnerability 29 and death. There is evidence of negative associations between greenspaces measured with 30 normalized difference vegetation index (NDVI) and frailty. However, NDVI is not as 31 informative as greenness structure indices which reflect characters such as shape and 32 connectivity. We aim to study the association between greenness structures and frailty in an 33 elderly Chinese cohort. 34 Methods: We included older adults from 2008-2014 waves of the China Longitudinal Healthy 35 Longevity Survey (CLHLS). We used greenspace indices from satellite to quantify greenspace 36 structures at county-level: area-edge, shape, and proximity, and calculated frailty index (FI) as 37 a health outcome. We did cross-sectional analyses using linear regression and logistical 38 regression, and longitudinal analyses using the generalized estimating equations (GEE). All 39 models were adjusted for covariates. 40 Results: Among 8,776 participants at baseline, the mean LPI, SHAPE, COHESION, and FI 41 were 7.93, 8.11, 97.6, and 0.17. The correlation between NDVI and greenness structure was 42 unnoticeable. In cross-sectional analyses, we found negative consistent dose-response 43 relationships for greenspace structures and frailty, especially in females, city residents, people 44 without a spouse, and with deteriorated frailty. Compared to participants living in the lowest 45 quartile of greenness structure, those in the highest quartile of LPI, SHAPE, and COHESION 46 had 32%, 44%, and 37% lower odds of frailty. However, we did not find a significant 47 association in longitudinal analyses due to higher mortality rate and FI of participants without 48 follow-up surveys. Conclusions: The larger value of area-edge, shape, and proximity is related to a lower 50 likelihood of frailty. Assessing complex shapes and connecting fragmentary greenspaces are 51 informative to public health through city planning.


55
Frailty refers to a geriatric syndrome that increases a person's vulnerability due to 56 degenerative changes and chronic diseases, reflecting cumulative physical, psychological, and 57 social deficits, which leads to higher risks of hospitalization, falls, depression, and mortality 1- reported that 7.0% of adults aged 60 years or older were frail 4 . 64 There is evidence of an association between greenness and frailty-related factors in 65 population health studies. A study in Hong Kong found that higher residential greenness levels 66 could improve frailty by mediating through physical activity, the number of diseases, and 67 cognitive functions 5 . A longitudinal study with 16,238 older adults with a 12-year follow-up in 68 China assessed greenery exposure at the neighborhood level, proving that higher residential 69 greenness levels are related to a lower likelihood of frailty, specifically in urban areas 6 . Besides, 70 mechanisms by which exposures to greenspaces promote healthy aging have been extensively 71 studied. First, more greenspaces in the residential environment could lead to fewer incidences 72 of loneliness, more social support, and improved social cohesion in the neighbourhood 7 . 73 Second, greenspaces may be a resource for psychological restoration 8 . Exposure to greenspaces 74 is associated with reduced stress and providing the opportunity to restore directed attention 9,10 , 75 which may benefit cognitive aging. Third, older adults living in areas with higher access to 76 greenspaces do more physical activities, which play a significant role in maintaining Moreover, we used NDVI from the Moderate Resolution Imaging Spectro-Radiometer 126 (MODIS) based on the longitude and latitude of each residential address as a measure of 127 greenness surrounding the residence 24 . NDVI ranges from −1.0 to 1.0, with larger values 128 indicating higher levels of vegetative density 25 . We deleted negative values, which represented 129 blue space or water. We calculated contemporaneous NDVI at the individual's residential 130 address at the death date for individuals who had died /the last interview date for those who 131 were alive and those lost to follow-up. The correlation between NDVI and greenness structure 132 was not noticeable. Therefore, the relationship between greenspaces and frailty can be more 133 accurately described using separate greenspace structures than using general NDVI values.  136 We used the Frailty Index (FI) to measure frailty status as the previous study 6,26 . FI is based 137 on the accumulation of aging-induced deficits, which is defined as the ratio of the number of 138 deficits existing in an individual divided by the total amount considered 27 . FI included 38 self-139 reported items, including instrumental activities of daily living, functional limitations, 140 activities of daily living, cognitive function, self-reported health status, interviewer-rated health 141 status, mental health, auditory and visual ability, heart rhythm, and chronic diseases (table S2).

Assessment of Frailty
We scored each term as 0 (absence of deficit) or 1 (presence of deficit) for 38 of 39 terms. We 143 scored the other 1 term as 2 if the participants reported 2 or more severe illnesses that caused 144 hospitalization or being bedridden in the past 2 years, such as stroke, cancer, and cataract. FI 145 was equal to the number of reported deficits divided by the total number of included deficits.

146
FI was a continuous variable and ranged from 0 to 1. A higher value indicated poorer frailty. 147 We also classified the continuous FI into two statuses: non-frail (FI ≤0.21) and frail (FI > 148 0.21) 28 . Changes of FI were the difference in FI scores measured between the last survey and 149 the baseline, categorized as no change or decrease, and an increase.  Statistical analysis 161 We hypothesized that larger value of area-edge, shape, and proximity were protective 162 factors for Chinese seniors' frailty, and the strength of this protection varied among the 163 subgroups. First, we used Pearson correlation coefficient ≥0.7 as a criterion for excluding the 164 greenspace indices given their collinearity. Second, a cross-sectional analysis was conducted 165 using linear regression and logistic regression to assess the associations between residential 166 greenness and frailty at baseline, adjusted for the study entrant year, age, sex, marital status, 167 geographic region, urban or rural residential location, literacy, annual household income, BMI, 168 smoking status, alcohol consumption, exercise status, and three-year average PM2.5. The linear 169 regression was conducted to assess NDVI, greenspace structures, and continuous baseline FI 170 scores. We also used a logistic regression model to calculate the odds ratio (OR) and 95% CIs 171 to indicate associations between indices of greenspace structures and binary FI status.

172
Considering the nonlinearity followed the process reported in the recent study, the three indices    Table 2 presents the association between greenness structure indices and frailty. In the 205 adjusted linear regression at baseline, there was an association between a higher value of 206 greenness structure indices and better frailty condition after adjustment. We observed a 207 significant dose-response relationship in the quartiles group. Each 0.1-unit increase in LPI, 208 SHAPE, and COHESION was statistical significantly associated with a 0.026-point, 0.028-209 point, and 0.025-point lower FI score in the fourth quartile. In the adjusted logistic regression, 210 an increase in all greenness structure indices was associated with an OR less than 1 of frailty.  Table 3 reports the relationship between greenness structure indices and changes in FI.  The results of sensitivity analysis were basically consistent with the main model, 221 indicating the specific indices type did not bias the results (table S4) 39 , which could increase memory, attention 40 , and mental health 41 . 253 Third, exposure to air pollution has been linked to respiratory diseases, and may be contributory 254 to frailty 42 . Minimized fragmentation and increased the largest patch percentage of green 255 structure could lower the mortality of pneumonia and chronic lower respiratory diseases by and 256 the mediation effects through reducing air pollutants 43 .

257
The present study observed that greenspace indices' protective effects were more evident 258 on city residents, people who were unmarried and not living with a spouse, and with 259 deteriorated frailty. A study in the Netherlands reported the significant association between 260 greenspaces and different perceived general health among different levels of urbanization 44 .

261
China has witnessed rapid urbanization widening the gap of unequal landscaping plans in urban 262 and rural areas 45,46 . Another possible explanation for urban-rural differences is that the lower 263 socioeconomic status, high competing risk from communicable diseases, and a persistent lack 264 of universal health coverage in rural areas weaken the greenspaces' positive function 47 .

265
Additionally, it might also be due to the urban-rural difference in FI at baseline (urban 0.18 vs.  48 . Therefore, if greenspaces can provide widowed elderly people with another form of 271 positive emotions, it will fill the marriage gap. We also found that participants with increasing 272 frailty were more likely to be beneficial from large and complex greenspaces. If the area and 273 shape of greenness can indeed delay the frailty process, greenspace planning can become a tool for healthy aging. 275 We observed a gender difference in the association between greenness structure and frailty. 276 Females tended to live in bigger and more complex green areas, and benefit more from greening 277 patterns than males. This is not consistent with the previous study in Hong Kong and the UK,

298
A protective association of greenness was identified in this study. We found that larger 299 areas, more complex shapes, and greater proximity were associated with a lower likelihood of 300 frailty among Chinese older adults. We observed a stronger association among females,